def test_flipping(self):
# Check flip
dt = DataTransforms(self.d)
flip_lr = dt.flip(direction="lr")
flip_ud = dt.flip(direction="ud")
check_lr = np.fliplr(self.d)
check_ud = np.flipud(self.d)
assert np.allclose(flip_ud, check_ud)
assert np.allclose(flip_lr, check_lr)
def test_median_filter(self): #Check median filter from PIL import Image, ImageFilter dt = DataTransforms(self.d) filtered = dt.median_filter(size=3) image = Image.fromarray(self.d) image = image.filter(ImageFilter.MedianFilter(size=3)) check_filtered = np.array(image) assert np.allclose(check_filtered, filtered)
def test_scaling(self): from PIL import Image # Check Scales dt = DataTransforms(self.d) h = 150 w = 150 scale = Image.fromarray(self.d).resize((h, w)) check_scale = dt.scale(h, w) np.allclose(scale, check_scale)
def test_gaussian_noise(self):
# check gaussian noise
dt = DataTransforms(self.d)
np.random.seed(0)
random_noise = self.d
random_noise = random_noise + np.random.normal(
loc=0, scale=25.5, size=self.d.shape)
np.random.seed(0)
check_random_noise = dt.gaussian_noise(mean=0, std=25.5)
assert np.allclose(random_noise, check_random_noise)
def test_shift(self):
# Check shift
dt = DataTransforms(self.d)
height = 5
width = 5
if len(self.d.shape) == 2:
shift = scipy.ndimage.shift(self.d, [height, width])
if len(self.d.shape) == 3:
shift = scipy.ndimage.shift(self.d, [height, width, 0])
check_shift = dt.shift(width, height)
assert np.allclose(shift, check_shift)
def test_salt_pepper_noise(self):
# check salt and pepper noise
dt = DataTransforms(self.d)
np.random.seed(0)
prob = 0.05
random_noise = self.d
noise = np.random.random(size=self.d.shape)
random_noise[noise < (prob / 2)] = 0
random_noise[noise > (1 - prob / 2)] = 255
np.random.seed(0)
check_random_noise = dt.salt_pepper_noise(prob, salt=255, pepper=0)
assert np.allclose(random_noise, check_random_noise)
def test_center_crop(self):
# Check center crop
dt = DataTransforms(self.d)
x_crop = 50
y_crop = 50
crop = dt.center_crop(x_crop, y_crop)
y = self.d.shape[0]
x = self.d.shape[1]
x_start = x // 2 - (x_crop // 2)
y_start = y // 2 - (y_crop // 2)
check_crop = self.d[y_start:y_start + y_crop, x_start:x_start + x_crop]
assert np.allclose(check_crop, crop)
def test_rotation(self):
# Check rotation
dt = DataTransforms(self.d)
angles = [0, 5, 10, 90]
for ang in angles:
rotate = dt.rotate(ang)
check_rotate = scipy.ndimage.rotate(self.d, ang)
assert np.allclose(rotate, check_rotate)
# Some more test cases for flip
rotate = dt.rotate(-90)
check_rotate = scipy.ndimage.rotate(self.d, 270)
assert np.allclose(rotate, check_rotate)
def test_blurring(self):
# Check Blurring
dt = DataTransforms(self.d)
blurred = dt.gaussian_blur(sigma=1.5)
check_blur = scipy.ndimage.gaussian_filter(self.d, 1.5)
assert np.allclose(check_blur, blurred)
def test_convert2gray(self):
# Check convert2gray
dt = DataTransforms(self.d)
gray = dt.convert2gray()
check_gray = np.dot(self.d[..., :3], [0.2989, 0.5870, 0.1140])
assert np.allclose(check_gray, gray)
